MiniMax-H3 multi-GPU parallel inference for ComfyUI | 多卡并行加速节点:2-8 GPU Ulysses sequence parallel, bit-identical video+audio generation
-
Updated
Sep 28, 2026 - Python
MiniMax-H3 multi-GPU parallel inference for ComfyUI | 多卡并行加速节点:2-8 GPU Ulysses sequence parallel, bit-identical video+audio generation
[🔥 ICML2026 🔥 ]A training-free orchestration framework for building interactive omni-modal assistants by composing off-the-shelf modality experts, explicit LLM routing, text-centric cross-modal memory, and interruption-aware streaming interaction.
AI 视频配乐与漫剧音效编辑器:ElevenLabs AI Music、音效生成、多轨混音与人声避让,适合 AI 漫剧、动态漫画、短剧和剪映后期;中文教程,可本地运行与二次开发。
AI 漫剧与短视频自动混音:ElevenLabs Scribe + FFmpeg Auto Ducking,人声说话时自动压低 BGM,适合动态漫画、AI 视频、知识口播;中文教程与本地混音源码。
Free browser tool that evens out a video's audio - levels volume, gentle compression and EQ, peak limiting - then saves a new MP4. Runs entirely on your device; nothing is uploaded. Less ear fatigue - smooth, even sound for your YouTube viewers.
Fix one-sided video audio on Android. Convert stereo to mono, keep the left or right channel, and save a new MP4 offline.
To associate your repository with the video-audio topic, visit your repo's landing page and select "manage topics."